MOLECULAR AND CELLULAR BIOLOGY, Sept. 1992, p. 4215-4229
Vol. 12, No. 9
0270-7306/92/094215-15$02.00/0 Copyright © 1992, American Society for Microbiology
Identification of Pre-mRNA Polyadenylation Sites in Saccharomyces cerevisiae STEFAN HEIDMANN, BRIGITTE OBERMAIER, KARIN VOGEL,
AND
HORST DOMDEY*
Laboratorium ftir Molekulare Biologie-Genzentrum-der Ludwig-Maximilians-Universitat Munchen, Am Klopferspitz 18a, D-8033 Martinsrie4, Germany Received 19 December 1991/Returned for modification 27 February 1992/Accepted 12 June 1992
In contrast to higher eukaryotes, little is known about the nature of the sequences which direct 3'-end formation of pre-mRNAs in the yeast Saccharomyces cerevisiae. The hexanucleotide AAUAAA, which is highly conserved and crucial in mammals, does not seem to have any functional importance for 3'-end formation in yeast cells. Instead, other elements have been proposed to serve as signal sequences. We performed a detailed investigation of the yeast ACT), ADHI, CYCI, and YPTI cDNAs, which showed that the polyadenylation sites used in vivo can be scattered over a region spanning up to 200 nucleotides. It therefore seems very unlikely that a single signal sequence is responsible for the selection of all these polyadenylation sites. Our study also showed that in the large majority of mRNAs, polyadenylation starts directly before or after an adenosine residue and that 3'-end formation of ADH1 transcripts occurs preferentially at the sequence PyAAA. Site-directed mutagenesis of these sites in theADHI gene suggested that this PyAAA sequence is essential for polyadenylation site selection both in vitro and in vivo. Furthermore, the 3'-terminal regions of the yeast genes investigated here are characterized by their capacity to act as signals for 3'-end formation in vivo in either orientation.
Processing of mRNA precursors is an essential event in expression in eukaryotes. Generally speaking, the maturation of a primary transcript generated by RNA polymerase II can be divided into several distinct processing reactions: capping of the 5' terminus of the primary transcript; modification of some adenosine residues; excision of intron sequences via splicing; and cleavage of the primary transcript at its 3' end and then addition of a poly(A) tail. The signals which govern cleavage of an mRNA precursor at its 3' end and addition of the poly(A) tail have been well defined in mammals; in addition to some weakly conserved signals in the DNA region downstream of the polyadenylation site, there is a strongly conserved hexanucleotide sequence (AATAAA) located around 10 to 30 nucleotides upstream of the polyadenylation site (for a review, see reference 48). In contrast, the 3' ends of the genes transcribed by RNA polymerase II in the yeast Saccharomyces cerevisiae do not contain such a strongly conserved nucleotide sequence. Only a subset of the yeast genes sequenced so far contain the hexanucleotide sequence AATAAA in their 3' regions, and for ADH2, it was shown that mutation or deletion of that signal did not impair correct 3'-end formation in vivo (19). Instead, a number of other different nucleotide sequences have been proposed which might serve as signals for processing of pre-mRNA 3' ends, such as the octanucleotide 'TiTIT-''lATA and the tripartite sequence TAG ...TA(T)GT.. 1T'T' (15, 50). More recently, these two motifs have been slightly modified to TITITl'AT (20) and TAG... .TATGTA (36), and there is also evidence for participation of the signal TATATA or TATCTA in the process of transcription termination and/or 3'-end processing (36). The presence or absence of these signals and the capacity of the 3'-end regions of some yeast genes to govern 3'-end formation in either orientation led to the postulation of different classes of polyadenylation sites (20). Most of the
published work has dealt with the yeast iso-1-cytochrome c (CYC1) gene, in which a 38-bp deletion was reported to abolish transcription termination and 3'-end processing (6, 50). Closer investigation of the deleted sequence and of revertants that displayed normal levels of CYCl mRNA led to the formulation of the above-mentioned signal sequences (36, 50). The observation that these sequence motifs are necessary but not sufficient to direct 3'-end formation indicates the participation of additional regulatory elements which may be located at or near the actual polyadenylation site (19, 20, 36). In this study, we investigated in greater detail the actual polyadenylation sites in various yeast genes and found that the polyadenylation sites in a given yeast gene can be distributed over a wide sequence range. Sitedirected mutagenesis and deletions of the ADHJ polyadenylation sites suggested that, in combination with distinct 3'-end-processing signals, the sequence Py(A)n is a preferential site for polyadenylation.
gene
*
Corresponding author.
MATERIALS AND METHODS Bacterial and yeast strains and culture conditions. Escherichia coli XL1 Blue (recAl endAl gyrA96 thi-1 hsdR17 supE44 reLAl lac [F' proAB lacJi ZAM15 TnlO(tet)] (5) was used for cloning procedures. Competent E. coli cells were produced by the method of Hanahan (14). E. coli was grown on Luria-Bertani medium supplemented with the appropriate antibiotics. Yeast strain XJ 24-24a (AL4Ta adeb arg4-17 trpl-1 tyr7-1 aro7-1) was the mRNA source of the cDNA library contained in the XZAPII vector (Stratagene). The mRNA for the rapid amplification of cDNA ends (RACE) employing the polymerase chain reaction (PCR) analysis was isolated from transformed DH484 cells (MATa ade2-1 leu2-3 leu2-112 canl-100 trpS-48 ura4-11 lysl-l). The yeast whole-cell extracts were prepared from strain EJ101 (AL4Ta trpl prol-126 prbl-112 pep4-3 prcl-126). Untransformed yeast cells were grown on yeast extract-peptone-dextrose medium. S. cere4215
4216
HEIDMANN ET AL.
visiae DH484 was transformed by the lithium acetate procedure (21). Transformants were selected on minimal medium supplemented with all necessary constituents except for leucine. Enzymes and chemicals. Restriction endonucleases, T7 RNA polymerase, T4 DNA ligase, avian myeloblastosis virus reverse transcriptase, exonuclease III, and T4 polynucleotide kinase were obtained from Boehringer Mannheim. SP6 RNA polymerase, human placental RNase inhibitor,
[^y-32P]ATP (185 TBq/mmol), [a- 2P]dCTP (111 TBq/mmol), [a-32P]UTP (29.6 TBq/mmol), and a-35S-dATP (24 TBq/ mmol) were from Amersham. Exonuclease VII was obtained
from Bethesda Research Laboratories. Taq polymerase was purchased from Cetus Corporation. Si nuclease was from Pharmacia. The following oligonucleotides were synthesized with an automated DNA synthesizer from Applied Biosystems: 01: 02: 03: 04: 05: 06: 07: 08:
5'-AATAACCAAAGCAGCAACCT-3' 5'-CGACGTAACATAGTTTTTCC-3' 5'-TGGGGCTCTGAATCTTTC-3'
5'-GACTCGAGTCTAGAGAGCTCC(T)17-3' 5'-GACTCGAGTCTAGAGAGCTCC-3' 5'-CAAGAATTCACCAACTGGGACG-3' 5'-GATCTAAGCTTGCGAATTCTCTAGAGTCGACCTGCAGA-3'
5'-GATCTCTGCAGGTCGACTCTAGAGAATTCGCAAGCTTA-3' Reverse transcription and amplification of specific polyadenylated mRNAs by RACE-PCR. RACE-PCR was done by the method of Frohman et al. (9) with some modifications. Poly(A)+ RNA (2 j,g) was incubated at 42°C for 90 min in a 20-pl reaction mixture containing 50 mM KCl, 20 mM Tris Cl (pH 8.4), 6 mM MgCl2, 1 mM each deoxynucleoside triphosphate (dNTP), 1 mM dithiothreitol, 2.4 pmol of oligonucleotide 04, 7 U of RNAsin, and 25 U of avian myeloblastosis virus reverse transcriptase. After inactivation of the reverse transcriptase for 3 min at 95°C, the reaction mixture was brought to a volume of 1 ml with TE (10 mM Tris Cl [pH 7.5], 0.1 mM EDTA) to dilute oligonucleotide 04. A 5-pl sample of this cDNA preparation was used in a PCR amplification in 50 pl containing 50 mM KCl, 20 mM Tris Cl (pH 8.4), 6 mM MgCl2, 100 puM each dNTP, 5 pmol of both 05 and 06, and 1 U of Taq DNA polymerase. The reaction mixtures were overlayered with 50 pl of mineral oil and incubated in a Perkin-Elmer thermal cycler with the following parameters. For the first cycle, denaturation lasted for 3 min at 94°C, annealing for 3 min at 50°C, and extension for 40 s at 72°C (since the gene-specific upstream primer 06 with the 5'protruding EcoRI site did not exactly match the template, it was necessary to run the first amplification cycle at a lower annealing temperature). The subsequent 30 PCR cycles were performed with the following parameters: denaturation for 40 s at 94°C, annealing for 30 s at 60°C, and extension for 40 s at 72°C. A 15-ptl sample of the reaction mixtures was analyzed on an agarose gel; to 20 pul of the samples, 4 pl of a solution containing 2.5% sodium dodecyl sulfate, 130 mM EDTA, and 2 mg of proteinase K per ml was added, and the mixture was incubated for 45 min at 37°C to digest the Taq DNA polymerase. After extraction twice with phenol and once with chloroform, the PCR products were precipitated with ethanol and then digested with EcoRI and XbaI (sites present in the primers) and ligated into an EcoRI-XbaI-
digested pBluescriptKSII- vector (Stratagene). Plasmid constructions and manipulations. The 3'-end region of the yeast ADHI gene was isolated as a 463-bp BamHI-HindIII fragment from plasmid pAAH5 (2) and
MOL. CELL. BIOL.
cloned into a BamHI-HindIII-digested pGEM2 vector (Promega) to provide plasmid pgADH463, from which transcription with T7 RNA polymerase led to transcripts in sense orientation. The insert was shortened from its 3' end by treatment with exonucleases III and VII after digestion with BamHI and SacI by the method of Yanisch-Perron et al. (49). The insert was also shortened from its 5' end up to two nucleotides upstream of the ADH1 translational stop codon by in vitro mutagenesis. This fragment had been shown earlier to direct 3'-end formation efficiently in vivo (45). The combination of deletions from both ends, each of them still working in vivo, resulted in a fragment of 91 bp. The corresponding complementary oligonucleotides with additional 5'-protruding GATC ends were synthesized, hybridized, and cloned in both orientations into the BamHIcleaved vector pBluescriptIIKS- to yield the plasmids pKSADH91s and pKSADH91a, where s (sense) and a (antisense) refer to the orientation of the insert relative to the T7 promoter. The ACTI terminator region was obtained as a 354-bp Sau3AI fragment from the plasmid pYactI (32). This region was cloned into a BamHI-cleaved pGEM2 vector, so that transcription with SP6 RNA polymerase from this plasmid (pgACT) yielded RNAs in the sense orientation. The YPT1 terminator region was obtained as a 494-bp Sau3AI fragment from the plasmid pYactI (32) and was cloned similarly into the BamHI-digested vector pBluescript IIKS- in the sense orientation. The 243-bp CYCl terminator region was assembled with synthetic oligodeoxynucleotides. It corresponds to the 236-bp TaqI fragment used by Butler and Platt (6) with additional Sau3AI sites at both ends. This fragment was cloned into the BamHI-cleaved vector pGEM2 in the sense orientation relative to the T7 promoter. For in vivo analysis, all investigated fragments were inserted into a truncated ACTI transcription unit starting with the BamHI site upstream of the promoter and ending at the internal HindIII site with the ACTI intron and the internal 321-bp BglII fragment deleted (32). This transcription unit was contained as a 933-bp BamHI-HindIII fragment in the yeast-E. coli shuttle vector YEp351 (16), to yield the plasmid pSH50 (see Fig. 5). The 5' overhangs of the 463-bp HindIII-BamHI ADH1 terminator fragment were filled in with Klenow polymerase, and HindIII linkers were ligated to the ends. The product was cleaved with HindIII and cloned into the HindIII-digested pSH50 to yield the plasmids pSH200 and pSH200a with theADHl terminator in the sense and antisense orientations, respectively. The 354-bp ACTI, the 243-bp CYCI, and the 494-bp YPTI 3'-terminal regions were cloned as Sau3AI fragments in both orientations into the BglII site of pSH200. The ADHI deletion fragments were tested in a different construct. First, the 3'-terminal region of theACTi gene was cloned in the sense orientation, as a 386-bp SmaI-AluI fragment isolated from plasmid pgACT, into the Klenow filled-in HindIII site of pSH50 (see Fig. 5) to yield plasmid pSH60. A subsequent in vitro mutagenesis was necessary to destroy the EcoRI site in the coding region of the LEU2 gene contained in the vector. Therefore, the origin of replication of the single-stranded bacteriophage fl was cloned as a 529-bp SacI-BamHI fragment obtained from the plasmid pUCfl (Pharmacia) into the SacI-Bam HI-cleaved plasmid pSH60. The mutagenesis was performed by using the BioRad in vitro mutagenesis kit based on the method described by Kunkel (23), and the EcoRI site of the YEp351 polylinker was destroyed by Klenow fill-in and subsequent religation.
VOL. 12, 1992
The two complementary oligonucleotides 07 and 08 were cloned into the BglII site of the truncated actin transcription unit to yield the plasmid pSH100. The deletion clones of the ADH1 3'-end region could then easily be cloned in the sense orientation into pSH100 as EcoRI-HindIII fragments obtained from the pGEM plasmids. The DNA inserts were subcloned by standard procedures (25). DNA sequencing. The XZAPII cDNA clones were sequenced by the chain termination method of Sanger et al. (39) as modified for supercoiled plasmids (8). The cloned RACE-PCR products were sequenced with the 373A DNA Sequencer from Applied Biosystems, using fluorescencelabeled primers and the thermal cycling protocol according to the manufacturer's protocols. In vitro RNA synthesis and processing reactions. In vitro RNA synthesis was performed as described previously (28); in the reactions, 7mG(5')ppp(5')G was included in 10-fold excess compared with the concentration of GTP so that most of the transcripts were capped at their 5' ends. The products were separated by electrophoresis on 6% polyacrylamide-8 M urea gels. The excised and eluted RNAs were used in in vitro processing reactions as described by Butler and Platt (6). The yeast extracts were prepared either as described by Lin et al. (24) or following a modified protocol by Butler et al. (7). RNA analysis. Yeast total RNA was prepared by the hot-phenol method as described by Kohrer and Domdey (22). Poly(A)+ RNA was isolated with an mRNA purification kit purchased from Pharmacia. For Northern (RNA) blotting, 1.5 ,ug of glyoxylated poly(A)+ RNA was separated on a 1.5% agarose gel and transferred (46) to Hybond N nylon membranes (Amersham). The oligonucleotides 01, 02, and 03 served as hybridization probes, with 01 hybridizing immediately downstream of the actin intron (220 nucleotides upstream of the BglII site), 02 hybridizing 60 bp, and 03 hybridizing 179 bp downstream of the BglII site (see Fig. SB). Nuclease Si analysis. Nuclease Si mapping of the 3' ends of the mRNAs was done by the method of Berk and Sharp (4) with some modifications. The DNA probes were isolated as BglII-PstI fragments from plasmid pSH101 and its derivatives, and therefore their ends were not able to hybridize to genuine ADHI mRNA. The probes were end labeled with [a-32P]dCTP and the Klenow fragment of DNA polymerase I at the BglII site. The labeled DNA fragments (50,000 cpm each) were denatured by boiling for 2 min and quenched on ice, and 2 ,ug of poly(A)+ RNA and 3 ,ug of carrier tRNA were added. The mixture was ethanol precipitated and resuspended in 20 pl of hybridization solution containing 80% deionized formamide, 40 mM PIPES [piperazine-N,N'bis(2-ethanesulfonic acid)], 0.4 M NaCl, and 1 mM EDTA. The samples were heated for 10 min at 65°C and hybridized for 16 h at 30°C. A 300-,ul portion of an ice-cold solution containing 0.28 M NaCl, 50 mM sodium acetate (pH 5.0), 4.5 mM ZnSO4, 20 ,ug of single-stranded salmon sperm DNA per ml, and 600 U of nuclease S1 per ml was added. After 90 min of incubation at 16 or 23°C, reactions were stopped by adding 80 pJ of stop solution containing 3.6 M ammonium acetate, 20 mM EDTA, and 40 ,ug of tRNA per ml. The low temperatures were chosen to avoid double-strand cleavage by nuclease S1 in the base-paired A+T-rich regions of the probes. The samples were extracted with phenol-chloroform and chloroform and precipitated with ethanol. Reaction products were separated on a 6% polyacrylamide-8 M urea denaturing gel.
YEAST PRE-mRNA POLYADENYLATION SITES
4217
RESULTS Efficiency of 3'-end processing varies with different RNA substrates. As previously shown, in vitro transcripts spanning the 3'-end regions of a number of yeast genes can be accurately cleaved and polyadenylated in a yeast whole-cell extract (1, 6, 7, 38). To extend these analyses and to improve our understanding of the mechanism of 3'-end processing we subjected synthetic RNA transcripts representing the 3' ends of the yeast ACT1, ADHI, and YPTJ genes (Fig. 1A) to in vitro 3'-end-processing reactions. It seemed reasonable to examine in vitro 3'-end processing of the corresponding antisense transcripts, because for the ADHI and YPTJ genes, antisense transcripts had been described (20, 47). Furthermore, as will be shown below, the corresponding DNA fragments directed efficient 3'-end formation in either orientation in vivo. All these RNAs were subjected to in vitro 3'-end-processing reactions under conditions previously described (6) in poly(A) polymerase-proficient and -deficient yeast extracts, respectively, and the products were analyzed by polyacrylamide gel electrophoresis (Fig. 2). The poly(A) polymerase-deficient extract was especially useful for the analysis of very inefficient processing reactions, since the small amounts of processed products were not spread out over a rather broad size range as a result of polyadenylation. The results are summarized in Table 1. The 3'-end-processing efficiency of the ACTI sense transcript was extremely low (Fig. 2), and faint bands corresponding to processing products were observed only after incubation in a partially fractionated extract prepared by the method of Butler et al. (7). Only after several days of exposure were strong signals detected (data not shown), which corresponded roughly to the previously described polyadenylation sites (11). A distinctly higher processing efficiency was observed for the YPT1, ADHI (Fig. 2), and CYCl (data not shown) transcripts in the sense orientation. The derived cleavage positions are in good, although not perfect, accordance with the polyadenylation sites previously determined in vivo (3, 6, 10). Analysis of the corresponding antisense transcripts showed a rather heterogeneous pattern. Whereas the ACTI andADHI antisense transcripts yielded clearly visible cleavage and polyadenylation products, no such products could be observed when the antisense transcript of CYCI (data not shown) was subjected to a processing reaction. Possible processing products of YPTI did not emerge until after several days of exposure. These results were obtained with both polyadenylation-proficient and -deficient extracts. Processing of the ACTI antisense transcript yielded four 5'-cleavage products in a narrow molecular weight range (Fig. 2, lower part). These products were very efficiently polyadenylated with an approximately 70-nucleotide poly(A) tail, which matches the previously determined lengths of poly(A) tails in yeast transcripts (13). Furthermore, the weak signals approximately 10 nucleotides above the 5'-cleavage products might indicate a biphasic polyadenylation process as described in higher eukaryotes (42), suggesting a similar mechanism in yeast cells. It is unlikely that these signals represent further 5'-cleavage products, because then they should also occur among the reaction products from tACTla when incubated in the polyadenylation-deficient extract. Sequences downstream from ADHI polyadenylation sites contribute to efficiency of the in vitro 3'-end-processing reaction. A set of experiments was designed to find out which part of the 3'-terminal region of the yeast ADHI gene was required to direct efficient 3'-end processing and polyadeny-
4218
MOL. CELL. BIOL.
HEIDMANN ET AL.
A
UAA
tACT1 s
Wm
I
Vol. 12, No. 9
0270-7306/92/094215-15$02.00/0 Copyright © 1992, American Society for Microbiology
Identification of Pre-mRNA Polyadenylation Sites in Saccharomyces cerevisiae STEFAN HEIDMANN, BRIGITTE OBERMAIER, KARIN VOGEL,
AND
HORST DOMDEY*
Laboratorium ftir Molekulare Biologie-Genzentrum-der Ludwig-Maximilians-Universitat Munchen, Am Klopferspitz 18a, D-8033 Martinsrie4, Germany Received 19 December 1991/Returned for modification 27 February 1992/Accepted 12 June 1992
In contrast to higher eukaryotes, little is known about the nature of the sequences which direct 3'-end formation of pre-mRNAs in the yeast Saccharomyces cerevisiae. The hexanucleotide AAUAAA, which is highly conserved and crucial in mammals, does not seem to have any functional importance for 3'-end formation in yeast cells. Instead, other elements have been proposed to serve as signal sequences. We performed a detailed investigation of the yeast ACT), ADHI, CYCI, and YPTI cDNAs, which showed that the polyadenylation sites used in vivo can be scattered over a region spanning up to 200 nucleotides. It therefore seems very unlikely that a single signal sequence is responsible for the selection of all these polyadenylation sites. Our study also showed that in the large majority of mRNAs, polyadenylation starts directly before or after an adenosine residue and that 3'-end formation of ADH1 transcripts occurs preferentially at the sequence PyAAA. Site-directed mutagenesis of these sites in theADHI gene suggested that this PyAAA sequence is essential for polyadenylation site selection both in vitro and in vivo. Furthermore, the 3'-terminal regions of the yeast genes investigated here are characterized by their capacity to act as signals for 3'-end formation in vivo in either orientation.
Processing of mRNA precursors is an essential event in expression in eukaryotes. Generally speaking, the maturation of a primary transcript generated by RNA polymerase II can be divided into several distinct processing reactions: capping of the 5' terminus of the primary transcript; modification of some adenosine residues; excision of intron sequences via splicing; and cleavage of the primary transcript at its 3' end and then addition of a poly(A) tail. The signals which govern cleavage of an mRNA precursor at its 3' end and addition of the poly(A) tail have been well defined in mammals; in addition to some weakly conserved signals in the DNA region downstream of the polyadenylation site, there is a strongly conserved hexanucleotide sequence (AATAAA) located around 10 to 30 nucleotides upstream of the polyadenylation site (for a review, see reference 48). In contrast, the 3' ends of the genes transcribed by RNA polymerase II in the yeast Saccharomyces cerevisiae do not contain such a strongly conserved nucleotide sequence. Only a subset of the yeast genes sequenced so far contain the hexanucleotide sequence AATAAA in their 3' regions, and for ADH2, it was shown that mutation or deletion of that signal did not impair correct 3'-end formation in vivo (19). Instead, a number of other different nucleotide sequences have been proposed which might serve as signals for processing of pre-mRNA 3' ends, such as the octanucleotide 'TiTIT-''lATA and the tripartite sequence TAG ...TA(T)GT.. 1T'T' (15, 50). More recently, these two motifs have been slightly modified to TITITl'AT (20) and TAG... .TATGTA (36), and there is also evidence for participation of the signal TATATA or TATCTA in the process of transcription termination and/or 3'-end processing (36). The presence or absence of these signals and the capacity of the 3'-end regions of some yeast genes to govern 3'-end formation in either orientation led to the postulation of different classes of polyadenylation sites (20). Most of the
published work has dealt with the yeast iso-1-cytochrome c (CYC1) gene, in which a 38-bp deletion was reported to abolish transcription termination and 3'-end processing (6, 50). Closer investigation of the deleted sequence and of revertants that displayed normal levels of CYCl mRNA led to the formulation of the above-mentioned signal sequences (36, 50). The observation that these sequence motifs are necessary but not sufficient to direct 3'-end formation indicates the participation of additional regulatory elements which may be located at or near the actual polyadenylation site (19, 20, 36). In this study, we investigated in greater detail the actual polyadenylation sites in various yeast genes and found that the polyadenylation sites in a given yeast gene can be distributed over a wide sequence range. Sitedirected mutagenesis and deletions of the ADHJ polyadenylation sites suggested that, in combination with distinct 3'-end-processing signals, the sequence Py(A)n is a preferential site for polyadenylation.
gene
*
Corresponding author.
MATERIALS AND METHODS Bacterial and yeast strains and culture conditions. Escherichia coli XL1 Blue (recAl endAl gyrA96 thi-1 hsdR17 supE44 reLAl lac [F' proAB lacJi ZAM15 TnlO(tet)] (5) was used for cloning procedures. Competent E. coli cells were produced by the method of Hanahan (14). E. coli was grown on Luria-Bertani medium supplemented with the appropriate antibiotics. Yeast strain XJ 24-24a (AL4Ta adeb arg4-17 trpl-1 tyr7-1 aro7-1) was the mRNA source of the cDNA library contained in the XZAPII vector (Stratagene). The mRNA for the rapid amplification of cDNA ends (RACE) employing the polymerase chain reaction (PCR) analysis was isolated from transformed DH484 cells (MATa ade2-1 leu2-3 leu2-112 canl-100 trpS-48 ura4-11 lysl-l). The yeast whole-cell extracts were prepared from strain EJ101 (AL4Ta trpl prol-126 prbl-112 pep4-3 prcl-126). Untransformed yeast cells were grown on yeast extract-peptone-dextrose medium. S. cere4215
4216
HEIDMANN ET AL.
visiae DH484 was transformed by the lithium acetate procedure (21). Transformants were selected on minimal medium supplemented with all necessary constituents except for leucine. Enzymes and chemicals. Restriction endonucleases, T7 RNA polymerase, T4 DNA ligase, avian myeloblastosis virus reverse transcriptase, exonuclease III, and T4 polynucleotide kinase were obtained from Boehringer Mannheim. SP6 RNA polymerase, human placental RNase inhibitor,
[^y-32P]ATP (185 TBq/mmol), [a- 2P]dCTP (111 TBq/mmol), [a-32P]UTP (29.6 TBq/mmol), and a-35S-dATP (24 TBq/ mmol) were from Amersham. Exonuclease VII was obtained
from Bethesda Research Laboratories. Taq polymerase was purchased from Cetus Corporation. Si nuclease was from Pharmacia. The following oligonucleotides were synthesized with an automated DNA synthesizer from Applied Biosystems: 01: 02: 03: 04: 05: 06: 07: 08:
5'-AATAACCAAAGCAGCAACCT-3' 5'-CGACGTAACATAGTTTTTCC-3' 5'-TGGGGCTCTGAATCTTTC-3'
5'-GACTCGAGTCTAGAGAGCTCC(T)17-3' 5'-GACTCGAGTCTAGAGAGCTCC-3' 5'-CAAGAATTCACCAACTGGGACG-3' 5'-GATCTAAGCTTGCGAATTCTCTAGAGTCGACCTGCAGA-3'
5'-GATCTCTGCAGGTCGACTCTAGAGAATTCGCAAGCTTA-3' Reverse transcription and amplification of specific polyadenylated mRNAs by RACE-PCR. RACE-PCR was done by the method of Frohman et al. (9) with some modifications. Poly(A)+ RNA (2 j,g) was incubated at 42°C for 90 min in a 20-pl reaction mixture containing 50 mM KCl, 20 mM Tris Cl (pH 8.4), 6 mM MgCl2, 1 mM each deoxynucleoside triphosphate (dNTP), 1 mM dithiothreitol, 2.4 pmol of oligonucleotide 04, 7 U of RNAsin, and 25 U of avian myeloblastosis virus reverse transcriptase. After inactivation of the reverse transcriptase for 3 min at 95°C, the reaction mixture was brought to a volume of 1 ml with TE (10 mM Tris Cl [pH 7.5], 0.1 mM EDTA) to dilute oligonucleotide 04. A 5-pl sample of this cDNA preparation was used in a PCR amplification in 50 pl containing 50 mM KCl, 20 mM Tris Cl (pH 8.4), 6 mM MgCl2, 100 puM each dNTP, 5 pmol of both 05 and 06, and 1 U of Taq DNA polymerase. The reaction mixtures were overlayered with 50 pl of mineral oil and incubated in a Perkin-Elmer thermal cycler with the following parameters. For the first cycle, denaturation lasted for 3 min at 94°C, annealing for 3 min at 50°C, and extension for 40 s at 72°C (since the gene-specific upstream primer 06 with the 5'protruding EcoRI site did not exactly match the template, it was necessary to run the first amplification cycle at a lower annealing temperature). The subsequent 30 PCR cycles were performed with the following parameters: denaturation for 40 s at 94°C, annealing for 30 s at 60°C, and extension for 40 s at 72°C. A 15-ptl sample of the reaction mixtures was analyzed on an agarose gel; to 20 pul of the samples, 4 pl of a solution containing 2.5% sodium dodecyl sulfate, 130 mM EDTA, and 2 mg of proteinase K per ml was added, and the mixture was incubated for 45 min at 37°C to digest the Taq DNA polymerase. After extraction twice with phenol and once with chloroform, the PCR products were precipitated with ethanol and then digested with EcoRI and XbaI (sites present in the primers) and ligated into an EcoRI-XbaI-
digested pBluescriptKSII- vector (Stratagene). Plasmid constructions and manipulations. The 3'-end region of the yeast ADHI gene was isolated as a 463-bp BamHI-HindIII fragment from plasmid pAAH5 (2) and
MOL. CELL. BIOL.
cloned into a BamHI-HindIII-digested pGEM2 vector (Promega) to provide plasmid pgADH463, from which transcription with T7 RNA polymerase led to transcripts in sense orientation. The insert was shortened from its 3' end by treatment with exonucleases III and VII after digestion with BamHI and SacI by the method of Yanisch-Perron et al. (49). The insert was also shortened from its 5' end up to two nucleotides upstream of the ADH1 translational stop codon by in vitro mutagenesis. This fragment had been shown earlier to direct 3'-end formation efficiently in vivo (45). The combination of deletions from both ends, each of them still working in vivo, resulted in a fragment of 91 bp. The corresponding complementary oligonucleotides with additional 5'-protruding GATC ends were synthesized, hybridized, and cloned in both orientations into the BamHIcleaved vector pBluescriptIIKS- to yield the plasmids pKSADH91s and pKSADH91a, where s (sense) and a (antisense) refer to the orientation of the insert relative to the T7 promoter. The ACTI terminator region was obtained as a 354-bp Sau3AI fragment from the plasmid pYactI (32). This region was cloned into a BamHI-cleaved pGEM2 vector, so that transcription with SP6 RNA polymerase from this plasmid (pgACT) yielded RNAs in the sense orientation. The YPT1 terminator region was obtained as a 494-bp Sau3AI fragment from the plasmid pYactI (32) and was cloned similarly into the BamHI-digested vector pBluescript IIKS- in the sense orientation. The 243-bp CYCl terminator region was assembled with synthetic oligodeoxynucleotides. It corresponds to the 236-bp TaqI fragment used by Butler and Platt (6) with additional Sau3AI sites at both ends. This fragment was cloned into the BamHI-cleaved vector pGEM2 in the sense orientation relative to the T7 promoter. For in vivo analysis, all investigated fragments were inserted into a truncated ACTI transcription unit starting with the BamHI site upstream of the promoter and ending at the internal HindIII site with the ACTI intron and the internal 321-bp BglII fragment deleted (32). This transcription unit was contained as a 933-bp BamHI-HindIII fragment in the yeast-E. coli shuttle vector YEp351 (16), to yield the plasmid pSH50 (see Fig. 5). The 5' overhangs of the 463-bp HindIII-BamHI ADH1 terminator fragment were filled in with Klenow polymerase, and HindIII linkers were ligated to the ends. The product was cleaved with HindIII and cloned into the HindIII-digested pSH50 to yield the plasmids pSH200 and pSH200a with theADHl terminator in the sense and antisense orientations, respectively. The 354-bp ACTI, the 243-bp CYCI, and the 494-bp YPTI 3'-terminal regions were cloned as Sau3AI fragments in both orientations into the BglII site of pSH200. The ADHI deletion fragments were tested in a different construct. First, the 3'-terminal region of theACTi gene was cloned in the sense orientation, as a 386-bp SmaI-AluI fragment isolated from plasmid pgACT, into the Klenow filled-in HindIII site of pSH50 (see Fig. 5) to yield plasmid pSH60. A subsequent in vitro mutagenesis was necessary to destroy the EcoRI site in the coding region of the LEU2 gene contained in the vector. Therefore, the origin of replication of the single-stranded bacteriophage fl was cloned as a 529-bp SacI-BamHI fragment obtained from the plasmid pUCfl (Pharmacia) into the SacI-Bam HI-cleaved plasmid pSH60. The mutagenesis was performed by using the BioRad in vitro mutagenesis kit based on the method described by Kunkel (23), and the EcoRI site of the YEp351 polylinker was destroyed by Klenow fill-in and subsequent religation.
VOL. 12, 1992
The two complementary oligonucleotides 07 and 08 were cloned into the BglII site of the truncated actin transcription unit to yield the plasmid pSH100. The deletion clones of the ADH1 3'-end region could then easily be cloned in the sense orientation into pSH100 as EcoRI-HindIII fragments obtained from the pGEM plasmids. The DNA inserts were subcloned by standard procedures (25). DNA sequencing. The XZAPII cDNA clones were sequenced by the chain termination method of Sanger et al. (39) as modified for supercoiled plasmids (8). The cloned RACE-PCR products were sequenced with the 373A DNA Sequencer from Applied Biosystems, using fluorescencelabeled primers and the thermal cycling protocol according to the manufacturer's protocols. In vitro RNA synthesis and processing reactions. In vitro RNA synthesis was performed as described previously (28); in the reactions, 7mG(5')ppp(5')G was included in 10-fold excess compared with the concentration of GTP so that most of the transcripts were capped at their 5' ends. The products were separated by electrophoresis on 6% polyacrylamide-8 M urea gels. The excised and eluted RNAs were used in in vitro processing reactions as described by Butler and Platt (6). The yeast extracts were prepared either as described by Lin et al. (24) or following a modified protocol by Butler et al. (7). RNA analysis. Yeast total RNA was prepared by the hot-phenol method as described by Kohrer and Domdey (22). Poly(A)+ RNA was isolated with an mRNA purification kit purchased from Pharmacia. For Northern (RNA) blotting, 1.5 ,ug of glyoxylated poly(A)+ RNA was separated on a 1.5% agarose gel and transferred (46) to Hybond N nylon membranes (Amersham). The oligonucleotides 01, 02, and 03 served as hybridization probes, with 01 hybridizing immediately downstream of the actin intron (220 nucleotides upstream of the BglII site), 02 hybridizing 60 bp, and 03 hybridizing 179 bp downstream of the BglII site (see Fig. SB). Nuclease Si analysis. Nuclease Si mapping of the 3' ends of the mRNAs was done by the method of Berk and Sharp (4) with some modifications. The DNA probes were isolated as BglII-PstI fragments from plasmid pSH101 and its derivatives, and therefore their ends were not able to hybridize to genuine ADHI mRNA. The probes were end labeled with [a-32P]dCTP and the Klenow fragment of DNA polymerase I at the BglII site. The labeled DNA fragments (50,000 cpm each) were denatured by boiling for 2 min and quenched on ice, and 2 ,ug of poly(A)+ RNA and 3 ,ug of carrier tRNA were added. The mixture was ethanol precipitated and resuspended in 20 pl of hybridization solution containing 80% deionized formamide, 40 mM PIPES [piperazine-N,N'bis(2-ethanesulfonic acid)], 0.4 M NaCl, and 1 mM EDTA. The samples were heated for 10 min at 65°C and hybridized for 16 h at 30°C. A 300-,ul portion of an ice-cold solution containing 0.28 M NaCl, 50 mM sodium acetate (pH 5.0), 4.5 mM ZnSO4, 20 ,ug of single-stranded salmon sperm DNA per ml, and 600 U of nuclease S1 per ml was added. After 90 min of incubation at 16 or 23°C, reactions were stopped by adding 80 pJ of stop solution containing 3.6 M ammonium acetate, 20 mM EDTA, and 40 ,ug of tRNA per ml. The low temperatures were chosen to avoid double-strand cleavage by nuclease S1 in the base-paired A+T-rich regions of the probes. The samples were extracted with phenol-chloroform and chloroform and precipitated with ethanol. Reaction products were separated on a 6% polyacrylamide-8 M urea denaturing gel.
YEAST PRE-mRNA POLYADENYLATION SITES
4217
RESULTS Efficiency of 3'-end processing varies with different RNA substrates. As previously shown, in vitro transcripts spanning the 3'-end regions of a number of yeast genes can be accurately cleaved and polyadenylated in a yeast whole-cell extract (1, 6, 7, 38). To extend these analyses and to improve our understanding of the mechanism of 3'-end processing we subjected synthetic RNA transcripts representing the 3' ends of the yeast ACT1, ADHI, and YPTJ genes (Fig. 1A) to in vitro 3'-end-processing reactions. It seemed reasonable to examine in vitro 3'-end processing of the corresponding antisense transcripts, because for the ADHI and YPTJ genes, antisense transcripts had been described (20, 47). Furthermore, as will be shown below, the corresponding DNA fragments directed efficient 3'-end formation in either orientation in vivo. All these RNAs were subjected to in vitro 3'-end-processing reactions under conditions previously described (6) in poly(A) polymerase-proficient and -deficient yeast extracts, respectively, and the products were analyzed by polyacrylamide gel electrophoresis (Fig. 2). The poly(A) polymerase-deficient extract was especially useful for the analysis of very inefficient processing reactions, since the small amounts of processed products were not spread out over a rather broad size range as a result of polyadenylation. The results are summarized in Table 1. The 3'-end-processing efficiency of the ACTI sense transcript was extremely low (Fig. 2), and faint bands corresponding to processing products were observed only after incubation in a partially fractionated extract prepared by the method of Butler et al. (7). Only after several days of exposure were strong signals detected (data not shown), which corresponded roughly to the previously described polyadenylation sites (11). A distinctly higher processing efficiency was observed for the YPT1, ADHI (Fig. 2), and CYCl (data not shown) transcripts in the sense orientation. The derived cleavage positions are in good, although not perfect, accordance with the polyadenylation sites previously determined in vivo (3, 6, 10). Analysis of the corresponding antisense transcripts showed a rather heterogeneous pattern. Whereas the ACTI andADHI antisense transcripts yielded clearly visible cleavage and polyadenylation products, no such products could be observed when the antisense transcript of CYCI (data not shown) was subjected to a processing reaction. Possible processing products of YPTI did not emerge until after several days of exposure. These results were obtained with both polyadenylation-proficient and -deficient extracts. Processing of the ACTI antisense transcript yielded four 5'-cleavage products in a narrow molecular weight range (Fig. 2, lower part). These products were very efficiently polyadenylated with an approximately 70-nucleotide poly(A) tail, which matches the previously determined lengths of poly(A) tails in yeast transcripts (13). Furthermore, the weak signals approximately 10 nucleotides above the 5'-cleavage products might indicate a biphasic polyadenylation process as described in higher eukaryotes (42), suggesting a similar mechanism in yeast cells. It is unlikely that these signals represent further 5'-cleavage products, because then they should also occur among the reaction products from tACTla when incubated in the polyadenylation-deficient extract. Sequences downstream from ADHI polyadenylation sites contribute to efficiency of the in vitro 3'-end-processing reaction. A set of experiments was designed to find out which part of the 3'-terminal region of the yeast ADHI gene was required to direct efficient 3'-end processing and polyadeny-
4218
MOL. CELL. BIOL.
HEIDMANN ET AL.
A
UAA
tACT1 s
Wm
I
